key: cord-0725156-9tncfs3d authors: Vattiato, Giorgia; Maclaren, Oliver; Lustig, Audrey; Binny, Rachelle N.; Hendy, Shaun C.; Plank, Michael J. title: An assessment of the potential impact of the Omicron variant of SARS-CoV-2 in Aotearoa New Zealand date: 2022-04-09 journal: Infect Dis Model DOI: 10.1016/j.idm.2022.04.002 sha: ad28d351365e29de2b761aab810c901a64c8b403 doc_id: 725156 cord_uid: 9tncfs3d New Zealand delayed the introduction of the Omicron variant of SARS-CoV-2 into the community by the continued use of strict border controls through to January 2022. This allowed time for vaccination rates to increase and the roll out of third doses of the vaccine (boosters) to begin. It also meant more data on the characteristics of Omicron became available prior to the first cases of community transmission. Here we present a mathematical model of an Omicron epidemic, incorporating the effects of the booster roll out and waning of vaccine-induced immunity, and based on estimates of vaccine effectiveness and disease severity from international data. The model considers differing levels of immunity against infection, severe illness and death, and ignores waning of infection-induced immunity. This model was used to provide an assessment of the potential impact of an Omicron wave in the New Zealand population, which helped inform government preparedness and response. At the time the modelling was carried out, the date of introduction of Omicron into the New Zealand community was unknown. We therefore simulated outbreaks with different start dates, as well as investigating different levels of booster uptake. We found that an outbreak starting on 1 February or 1 March led to a lower health burden than an outbreak starting on 1 January because of increased booster coverage, particularly in older age groups. We also found that outbreaks starting later in the year led to worse health outcomes than an outbreak starting on 1 March. This is because waning immunity in older groups started to outweigh the increased protection from higher booster coverage in younger groups. For an outbreak starting on 1 February and with high booster uptake, the number of occupied hospital beds in the model peaked between 800 and 3,300 depending on assumed transmission rates. We conclude that combining an accelerated booster programme with public health measures to flatten the curve are key to avoid overwhelming the healthcare system. New Zealand delayed the introduction of the Omicron variant of SARS-CoV-2 into the community by the continued use of strict border controls through to January 2022. This allowed time for vaccination rates to increase and the roll out of third doses of the vaccine (boosters) to begin. It also meant more data on the characteristics of Omicron became available prior to the first cases of community transmission. Here we present a mathematical model of an Omicron epidemic, incorporating the effects of the booster roll out and waning of vaccine-induced immunity, and based on estimates of vaccine effectiveness and disease severity from international data. The model considers differing levels of immunity against infection, severe illness and death, and ignores waning of infection-induced immunity. This model was used to provide an assessment of the potential impact of an Omicron wave in the New Zealand population, which helped inform government preparedness and response. At the time the modelling was carried out, the date of introduction of Omicron into the New Zealand community was unknown. We therefore simulated outbreaks with different start dates, as well as investigating different levels of booster uptake. We found that an outbreak starting on 1 February or 1 March led to a lower health burden than an outbreak starting on 1 January because of increased booster coverage, particularly in older age groups. We also found that outbreaks starting later in the year led to worse health outcomes than an outbreak starting on 1 March. This is because waning immunity in older groups started to outweigh the increased protection from higher booster coverage in younger groups. For an outbreak starting on 1 February and with high booster uptake, the number of occupied hospital beds in the model peaked between 800 and 3,300 depending on assumed transmission rates. We Since the early stages of the COVID-19 pandemic, New Zealand has used strong border controls to minimise the importation of SARS-CoV-2 into the community from overseas (1) . In August 2021, New Zealand experienced a community outbreak of the B.1.617.2 (Delta) variant. The outbreak was linked by whole genome sequencing to an ongoing Delta outbreak in the Australian state of New South Wales, and was likely caused by the virus leaking out of a managed isolation and quarantine facility. The outbreak was suppressed with a combination of non-pharmaceutical interventions, increasing levels of vaccination coverage, and ongoing use of managed isolation and quarantine for all international arrivals. By the end of 2021, New Zealand had achieved high vaccination coverage with over 90% of the eligible population having received two doses of the Pfizer/BioNTech BNT162b2 vaccine. The population had very low levels of prior infection, with around 12,600 total community cases since February 2020 out of a population of 5.1 million (0.25%). The health burden of the Delta outbreak was disproportionately felt by Māori and Pacific people (2) , groups known to be at higher risk from COVID-19 (3, 4) . On 26 November 2021, the World Health Organisation designated the B.1.1.529 variant of SARS-CoV-2 a variant of concern, named Omicron (5). This followed a rapid growth in COVID- 19 Africa driven by the emergence of the new variant (6) . In many populations, Omicron has exhibited a significant transmission advantage over Delta, at least partly due to a greater ability to infect people who have been vaccinated or previously infected (7) . At the time of Omicron's emergence, New Zealand still had strict border controls in place, with all international arrivals required to spend 7 days in government-managed isolation and quarantine. Border restrictions were scheduled to be relaxed in stages, starting from mid January 2022. However, as a result of the emergence and rapid spread of Omicron, these reopening plans were postponed. This delayed the introduction of Omicron into the community, with the first known cases of community transmission reported on 23 January 2022 (8) . By this time, the 7-day rolling average number of cases associated with the Delta outbreak had reduced to approximately 25 per day with less than 10 cases in hospital. Delaying the arrival of Omicron into New Zealand bought valuable time to increase vaccine coverage, start the paediatric vaccination programme, which began on 17 January 2022, and roll out booster doses. By 23 January 2022, approximately 77% of New Zealand's population (93% of the eligible population over 12 years old) had received at least two doses of the Pfizer-BioNTech BNT162b2 vaccine, and 19% had received a third dose (8) , referred to hereafter as a booster. Vaccination of children aged 5-11 years began on 17 January 2022. Booster coverage was higher in older age groups, who were prioritised in the initial vaccination programme and therefore became eligible for the booster earlier. However, Māori tended to receive their primary vaccine course later than non-Māori due to an inequitable vaccine rollout (9, 10) . Hence, given the required four-month interval after the second dose, a low proportion of Māori were eligible for the booster by early 2022. Given the very large number of cases among international arrivals in managed isolation and quarantine, and the high transmissibility of Omicron (6) , it was apparent by early January 2022 that it was a matter of time until a border-related outbreak sparked a major epidemic wave in New Zealand. Although experiences of Omicron in other countries serve as useful guides, New Zealand was in a different situation to most other jurisdictions, with a high national vaccination rate, low but rapidly rising levels of booster coverage, and negligible infection-acquired immunity. This paper presents a mathematical model that was used prior to the arrival of Omicron into the New Zealand community to assess the potential impact of an Omicron wave. The model builds on a previously published age-structured stochastic model for the Delta variant (11, 12) , with modified parameters for disease severity and vaccine effectiveness based on international data for Omicron, and generalised to include booster doses and waning of vaccine-induced immunity over time. The model also includes the effects of changing levels of two-dose and three-dose vaccine coverage in different age groups, based on a combination of real data on vaccines administered and projected future uptake. The results were used to inform government strategy and preparedness in January 2022, including likely demand on the healthcare system and the importance of achieving higher booster coverage. At the time the research was carried out, the date of introduction of Omicron into the New Zealand community was unknown. We therefore explored outcomes for a range of different outbreak start dates and different levels of booster coverage. This allowed the model to explore the interaction between increasing immunity due to the ongoing booster rollout and decreasing immunity due to waning. We use an age-structured stochastic model for transmission of SARS-CoV-2. This builds on a model that has previously been used to describe New Zealand's Delta outbreak that started from a borderrelated source in August 2021. With a time-varying control function fitted to data on new daily cases, the model provides a reasonably good fit to cases, hospitalisation and deaths using parameter values for the Delta variant (12) . Here we generalise the model to include the effects of booster doses and waning of vaccine-induced immunity. We also modify key parameter values to reflect the characteristics of the Omicron variant, as described below. For a complete model specification, see Supplementary Information. Software to run the model is available at https://github.com/michaelplanknz/assessmentof-potential-impact-of-omicron-in-NZ. Vaccine coverage by 5-year age band is as per data on doses administered up to 17 January 2022, with the additional assumption that everyone who had received their first dose by 17 January receives their J o u r n a l P r e -p r o o f second dose five weeks later. This means that approximately 91% of over-12-year olds are doublevaccinated. Note that model vaccine coverage may be lower than official Ministry of Health statistics because we used the StatsNZ estimated resident population (ERP) as population denominators (see Supplementary Table S1 ), rather than the health service utilisation (HSU) population, which is typically smaller. As of January 2021, adults become eligible for the booster dose 120 days after receiving their second dose and we assume that either 70% or 90% of adults receive their third dose two weeks after becoming eligible. Vaccination of 5-11-year-olds began on 17 January 2022, with an eight-week interval between the first and second dose. As a simple model of the effects of vaccinating this age group, we assume that there is a 75% uptake over an eight-week period. All vaccine doses are assumed to take effect 14 days after being received. doses, and three doses (under the assumption that 90% of those eligible get a booster dose), plotted by date the immunity takes effect, which is assumed to be 14 days after the vaccine is received. Vaccine effectiveness of the Pfizer/BioNTech BNT162b2 vaccine is characterised by three model parameters: reduction in risk of infection, risk of hospitalisation, and risk of death. Effectiveness is assumed to wane with time since most recent dose (Table 1) according to estimates by UKHSA (13, 14) . In addition, we assume that effectiveness against infection is the same as effectiveness against symptoms (which is supported by UK data from routine testing of healthcare workers (13)), but that there is no additional reduction in transmission for breakthrough infections. This is an optimistic assumption as regards infection prevention, but a pessimistic assumption as regards breakthrough transmission. We assume that effectiveness against death is halfway between effectiveness against hospitalisation and 100%, as vaccines tend to provide better protection against more severe outcomes. These assumptions are broadly consistent with the range of vaccine effectiveness values used by UK J o u r n a l P r e -p r o o f SPI-MO modelling groups (15) (16) (17) . In a sensitivity analysis, we investigate an alternative set of vaccine effectiveness parameters from Golding & Lydeamore (18) -see Supplementary Table S2 . There is significant uncertainty about the relative contribution of intrinsic transmissibility (as measured by 0 ), generation time, and immune evasion to Omicron's transmission advantage over Delta. (7) Here we investigate a baseline scenario in which the reproduction number excluding the effects of immunity The assumed risk of hospitalisation and death in five-year age bands for infections in unvaccinated people is shown in Table 2 . The risk of hospitalisation is based on the estimates of (24), adjusted as in our previous model by an odds ratio of 2.26 for the Delta variant (25) , and additionally adjusted by a hazard ratio of 0.33 reflecting lower intrinsic severity of Omicron relative to Delta (26-28). The risk of death is based on estimates of (24) adjusted by a hazard ratio of 0.3 for Omicron (29) . (27) found that the risk of death for Omicron cases was 0.19 times the risk of death for Delta cases. However, they did not have sufficient data to adjust this estimate for other covariates. Not controlling for vaccination status and prior infection means this may be an underestimate of the relative risk because Omicron cases are more likely to be breakthrough infections, which tend to be milder. Note that these hazard ratios describe the intrinsic severity of Omicron relative to Delta. The realised severity is the product of intrinsic severity with vaccine effectiveness against infection and hospitalisation, and the age and vaccination status of the subpopulation that becomes infected. The reduction in realised severity relative to Delta is a model output and may be more than, similar to, or less than the reduction in intrinsic severity (30). We assume that the average length of hospital stay for Omicron is 4 days (31), which is shorter than estimates for the Delta variant. Other simplifying assumptions Immunity from infections that occurred prior to the start of the simulated time period is ignored. This assumption is not likely to have a large effect on model results given that, prior to the arrival of Omicron in the community, New Zealand has had approximately 12,600 confirmed community cases of COVID-19, which is around 0.25% of the total population. We do not consider the effects of a concurrent outbreak of the Delta variant. This is reasonable given that the number of Delta cases had declined to a 7-day average of around 25 per day. Infection with the Omicron variant is assumed to provide complete protection against re-infection with Omicron for the remainder of the simulation. Differences in the effectiveness of the AstraZeneca vaccine relative to the Pfizer vaccine are ignored. This is expected to have a negligible effect on population-level outcomes as the number of AstraZeneca vaccines given in New Zealand is very small. Vaccine effectiveness and waning immunity for people who have had one vaccine dose is ignored. This affects a relatively small part of the population. The effects of seasonality are not included in the model. Simulations are initialised with 500 seed infections introduced over a one-week time period following the specified outbreak start date. This models some initial undetected community transmission and means that model outputs are restricted to seeding events that lead to established community transmission and exclude those that go stochastically extinct. Key model outputs are not highly sensitive to the number of seed infections, though it will affect the timing of the peak. J o u r n a l P r e -p r o o f Table 2 . Hospitalisation and death rates for unvaccinated infected people in five-year age bands. Table 4 caption for details of data sources). Across all scenarios investigated, an outbreak starting on 1 February leads to lower peaks and fewer cumulative cases, hospitalisations and deaths than an outbreak starting on 1 January. This is because of the increased level of booster coverage achieved by February/March (Figure 1 ). An outbreak starting on 1 March leads to similar outcomes to an outbreak staring on 1 February. Outbreaks starting later (1 April or 1 May) lead to progressively higher peaks and more cumulative cases, hospitalisations and deaths than an outbreak starting on 1 March. This is because the effects of waning immunity, particularly in older age groups who tend to be vaccinated and boosted earlier, start to outweigh the benefits from boosting diminishing number of people in younger groups. Finally, we investigated the sensitivity of the model to parameters affecting the age distribution of infections. The realised age distribution of cases for the Omicron wave in New Zealand was not known a priori and is a product of contact rates between age groups, two-dose and three-dose vaccine coverage by age (which varied through time), and vaccine effectiveness against infection and transmission (which are different for Omicron relative to Delta). We performed model simulations using an adjusted contact matrix with higher contact rates among younger groups, renormalised to give the same as in the baseline scenario (see Supplementary Information) . This adjusted contact matrix results in a younger age distribution of cases, which is closer to that observed in New Zealand's Delta outbreak and in Queensland's Omicron wave up to mid January 2022 ( Figure 5 ). Due to the shift of cases to younger groups, the number of hospitalisations and deaths (Table 5) Table 5 . Sensitivity analysis comparing model outputs for the baseline scenarios with alternative scenarios: (1) with no booster coverage; (2) using alternative vaccine effectiveness parameters (see Supplementary Table S2 ); (3) with an adjusted contact matrix modelling increased transmission among younger groups and decreased transmission among older groups relative to the baseline scenario. We have presented a preliminary assessment of the potential impact of the Omicron variant of SARS-CoV-2 in New Zealand. This modelling was carried out before the detection of the Omicron outbreak on 23 January 2022 and simulates outbreaks with a range of different hypothetical start dates. These results should therefore be treated as preliminary estimates of potential outcomes, and the relative effect of the booster rollout over time. Work is ongoing to model the unfolding epidemic wave in real time. For an outbreak starting around 1 February, in scenarios where there is high booster uptake, peak hospital admissions range from 200 to 800 per day, and peak demand for hospital beds ranges from 800 to 3,300 depending on assumed transmission rates. These numbers would put significant strain on hospital capacity, suggesting that public health measures aimed at flattening the curve may be necessary to avoid overwhelming the healthcare system. The Ministry of Health has reportedly said there are around 7,500 inpatient/ward beds in public hospitals, of which around 5,600 were occupied with non-COVID-19 cases in early March 2022 (33) . For comparison, the peak hospital occupancy during the Delta outbreak in 2021 was 93 (34). Note that limiting heath impacts to the lower end of the modelled ranges (i.e. the low transmission scenario) would likely only be possible with an effective public health response. Due to the effects of waning immunity and the ongoing booster and paediatric vaccine rollouts, different groups will have different levels of risk at different times. Groups that are not yet eligible for the booster will be at elevated risk of severe illness. This disproportionately includes Māori, who were not adequately prioritised due to an inequitable vaccine roll out (9), and who have higher risk of severe illness from COVID-19 (3) (4). In general, lower booster uptake leads to worse outcomes. Slowing an outbreak to allow time for more people to receive their booster dose is a strategy that could reduce the overall health burden. However, waning immunity presents a danger that health outcomes in groups that were earliest to receive their booster become worse after a longer time period. Outbreaks that occur after significant waning of immunity can result in a higher overall health burden compared to outbreaks that occur during peak immunity. Maintaining high immunity levels across all groups will be important for future vaccination strategies. Given the significant uncertainty about model parameters, we benchmarked model outputs against international data for selected jurisdictions. Each of these jurisdictions had their own particular characteristics including age structure, levels of vaccination and immunity from prior infection, prevalence of comorbidities, testing rates, behavioural change in response to perceived risk, mask use, Because of the strong age gradient in severity of COVID-19, the number of hospitalisations and deaths is sensitive to the age distribution of infections. The baseline model scenario uses a contact matrix defining contact rates between different age groups based on pre-pandemic survey data (35) and adjusted for the New Zealand population (11) . However, people in older, more at-risk age groups might become more cautious and reduce their contacts during an Omicron outbreak. We investigated the possible consequences of this using a modified contact matrix that assumed increased contact rates among younger groups relative to older groups. This resulted in a large decrease in hospitalisations and deaths compared to the baseline scenario. This is due to a shift in the age structure of infections towards younger groups, who have a lower risk of severe illness or death from COVID-19. These results underscore the importance of measures to reduce infection rates in older groups. Evidence about vaccine effectiveness and severity for Omicron relative to Delta is still accumulating and estimates of hospitalisation and death rates are subject to uncertainty. We have assumed that the observed prevention of symptomatic COVID-19 is due to infection prevention. However, if overall reduction in transmission is lower than this, our results for the number of cases could be significant underestimates. We assumed that vaccine effectiveness against death is higher than vaccine effectiveness against hospitalisation. This is consistent with observed patterns that vaccines tend to be more effective against more severe health outcomes than against milder outcomes. However, if vaccine effectiveness against death is closer to vaccine effectiveness against hospitalisation, the number of deaths could be up to double our estimates. Deaths are a particularly uncertain model output because, at the time the modelling was undertaken, there was limited direct data on vaccine effectiveness against death and the risk of death for the Omicron variant. We assumed that vaccine effectiveness is the same in all age groups. There is some evidence that the effectiveness of the Pfizer vaccine against symptomatic disease caused by Omicron wanes faster in over 65-year-olds than in younger groups (36). Although vaccine effectiveness against hospitalisation was still high, if waning immunity results in an increasing number of infections in older age groups, our results for the total number of hospitalisation and deaths could be underestimates. The generation interval of Omicron is a key model parameter that is uncertain at present. We have considered scenarios that range from a relatively high reproduction number and a generation interval that is similar to the ancestral strain of SARS-CoV-2 to a smaller reproduction number and shorter generation interval (19, 23 There could also be a long tail or a second wave following the initial peak if mixing drops and then increases again as normal behaviour resumes. The model assumes a relatively stable proportion of infections are reported as cases and does not take into account the effects of limited testing capacity. If testing capacity is exceeded, reported case numbers may cease to be a useful reflection of the true incidence of COVID-19. The prevalence of Delta in New Zealand prior to the start of the Omicron wave was very low. However, if the number of Delta cases increases greatly at the same time as an Omicron wave, this could significantly add to the health burden of the epidemic. It would also complicate situational awareness given the large number of anticipated cases, limited genome sequencing capacity, and differential risks of clinical outcomes for the two variants. It is possible that Delta could account for a low proportion of cases but a higher proportion of healthcare demand. In New Zealand's 2021 Delta outbreak, the proportion of cases hospitalised exceeded model predictions. This could be because the outbreak was concentrated in relatively high-risk groups. If this pattern is repeated in an Omicron outbreak, our results for hospitalisations could be underestimates. The The wave is still in progress and the cumulative number of cases, hospitalisations and deaths will increase. In addition, officially reported numbers of hospitalisations and deaths include everyone within 14 days or 28 days of a positive COVID -19 test respectively. As in other jurisdictions, these numbers will include a substantial proportion of incidental admissions and deaths, i.e. those whose primary cause of illness or death is not COVID-19. More detailed data on cause of admission or death is being collated but is not yet available. Therefore, a comparison with model results can only be partial at this stage. Overall, outcomes have fallen within the broad ranges estimated by the model scenarios. Uptake of boosters was strong in February, accelerated by a reduction in the required interval between the second and third dose from 4 months to 3 months. However, booster uptake slowed in March and is currently closer to the scenario with 70% uptake than to the scenario with 90% uptake. Reported cases have been reasonably consistent with the medium transmission scenario (with an outbreak start date of 1 February) in Table 3 , but hospitalisations and deaths have been closer to the low transmission scenario. This is at least partly explained by the fact that the age distribution of cases has been closer to that with the adjusted contract matrix (red bars in Figure 5 ) than with the original contact matrix (blue bars in Figure 5 ). Comparing with the "adjusted contact matrix" sensitivity analysis in Table 5 , the peak in cases was 50% higher than the model estimate and peak hospital occupancy was 2.6% higher. Deaths are already J o u r n a l P r e -p r o o f 19% higher than the projected total in this scenario. However, the likely inclusion of some incidental hospitalisations and deaths will affect these comparisons. Successful Elimination of Covid-19 Transmission in New Zealand Inequities and perspectives from the COVID-Delta outbreak: The imperative for strengthening the Pacific nursing workforce in Aotearoa New Zealand COVID-19: we must not forget about Indigenous health and equity Māori and Pacific People in New Zealand have higher risk of hospitalisation for COVID-19 Classification of Omicron (B.1.1.529): SARS-CoV-2 Variant of Concern Rapid epidemic expansion of the SARS-CoV-2 Omicron variant in southern Africa Bounding the levels of transmissibility and immune evasion of the Omicron variant in South Africa Haumaru: The Covid-19 Priority Report. Waitangi Tribunal Will access to COVID-19 vaccine in Aotearoa be equitable for priority populations? A COVID-19 vaccination model for Aotearoa New Zealand United Kingdom Health Security Agency. SARS-CoV-2 variants of concern and variants under investigation in England: Technical briefing 34 Vaccine Effectiveness against the Omicron (B.1.1.529) Variant Short-term Projections based on Early Omicron Variant Dynamics in England Projected epidemiological consequences of the Omicron SARS-CoV-2 variant in England Warwick Omicron Modelling. Scientific Advisory Group for Emergencies Analyses to predict the efficacy and waning of vaccines and previous infection against transmission and clinical outcomes of SARS-CoV-2 variants 2022 Estimation of the test to test distribution as a proxy for generation interval distribution for the Omicron variant in England. medRxiv Quantifying SARS-CoV-2 transmission suggests epidemic control with digital contact tracing Estimation of Serial Interval and Reproduction Number to Quantify the Transmissibility of SARS-CoV-2 Omicron Variant in South Korea Age-specific rate of severe and critical SARS-CoV-2 infections estimated with multi-country seroprevalence studies Hospital admission and emergency care attendance risk for SARS-CoV-2 delta (B. 1.617. 2) compared with alpha (B. 1.1. 7) variants of concern: a cohort study Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study Risk of COVID-19 related deaths for SARS-CoV-2 Omicron (B.1.1.529) compared with Delta (B.1.617.2). medRxiv. 2022:2022.02.24.22271466. 30. SPI-M-O Chairs. Statement on COVID-19 Decreased severity of disease during the first global Omicron variant Covid-19 outbreak in a large hospital in Tshwane, South Africa A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker) New Zealand's hospital and ICU beds in numbers, as Omicron cases rise Ministry of Health. Media release: 29 United Kingdom Health Security Agency. Effectiveness of 3 doses of COVID-19 vaccines against symptomatic COVID-19 and hospitalisation in adults aged 65 years and older United Kingdom Health Security Agency. SARS-CoV-2 variants of concern and variants under investigationin England: Technical briefing 37 CoV-Spectrum: analysis of globally shared SARS-CoV-2 data to identify and characterize new variants The authors acknowledge the support of the New Zealand Ministry of Health, StatsNZ, and the Institute J o u r n a l P r e -p r o o f ☐ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.☒The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:The authors were funded by the New Zealand Government to carry out this research and to provide advice via the cross-agency COVID-19 Modelling Steering Group.